JP2008044818A - Group iii-v nitride-based semiconductor substrate and group iii-v nitride-based light-emitting element - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a group III-V nitride-based semiconductor substrate which can obtain performances equivalent to the values of averaged electric characteristics, obtained by the measurement of a normal bulk crystal, and to provide a group III-V nitride-based light-emitting element. <P>SOLUTION: In the group III-V nitride-based semiconductor substrate, the minimum value of the concentration of a n-type impurity at an arbitrary place within the substrate surface is set to be ≥5×10<SP>17</SP>cm<SP>-3</SP>. Further, the group III-V nitride-based light-emitting element comprises the group III-V nitride-based semiconductor substrate and at least an active layer composed of a group III-V nitride-based semiconductor, which layer is formed on the group III-V nitride-based semiconductor substrate. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a group III-V nitride semiconductor substrate capable of avoiding problems such as electrode ohmic defects, device characteristic defects, and reliability defects, and group III-V nitride light emission using the semiconductor substrate. It relates to an element.

  Group III-V nitride-based semiconductor materials such as gallium nitride (GaN), indium gallium nitride (InGaN), and gallium aluminum nitride (GaAlN) have a sufficiently large forbidden band and the interband transition is also a direct transition type. Application to short-wavelength light emitting devices has been actively studied. In addition, application to electronic devices is also expected due to the high saturation drift velocity of electrons and the use of two-dimensional carrier gas by heterojunction.

  When a semiconductor device is manufactured, so-called homoepitaxial growth is generally performed using a substrate having the same lattice constant and linear expansion coefficient as that of an epitaxially grown crystal. For example, a GaAs single crystal substrate is used as a substrate for epitaxial growth of GaAs or AlGaAs. However, it has been impossible to produce a group III-V nitride semiconductor substrate having a size and characteristics sufficient for practical use so far only for a group III-V nitride semiconductor crystal such as GaN. For this reason, single crystal sapphire is used as a substrate for GaN growth, and metal organic vapor phase epitaxy (MOVPE), molecular beam vapor phase epitaxy (MBE), hydride vapor phase epitaxy (on the single crystal sapphire substrate ( GaN was heteroepitaxially grown by a vapor phase growth method such as HVPE.

  However, since the sapphire substrate has a lattice constant different from that of GaN, a single crystal film cannot be grown by directly growing GaN on the sapphire substrate. Therefore, by applying the low temperature buffer layer technology devised for hetero-growth of Si and the like on the sapphire substrate to the growth of GaN, a buffer layer of AlN or GaN is once formed on the sapphire substrate at a low temperature of about 500 ° C. A method has been devised in which GaN is grown on the low-temperature growth buffer layer after the growth and relaxation of lattice distortion.

  By using this low-temperature grown nitride layer as a buffer layer, single crystal epitaxial growth of GaN on a sapphire substrate became possible. However, the low-temperature buffer layer has a narrow optimum condition setting range for the growth temperature and the growth film thickness, and it is difficult to achieve reproducibility.

  For this reason, the device structure is not epitaxially grown continuously on the sapphire substrate, but only a substrate on which only a GaN single layer is grown, that is, a so-called template is collectively collected or only the GaN template is externally deposited. A method is also used in which a device structure is epitaxially grown on the device.

However, in the template in which GaN is grown on sapphire, even if the low-temperature buffer layer technique is used, the deviation of the lattice between the substrate and the crystal is still difficult, and the obtained GaN has a density of 10 ⁹ to 10 ¹⁰ cm ^−2. Have dislocation density. This defect is an obstacle to the production of GaN-based devices, especially LDs and ultraviolet light-emitting LEDs, so GaN templates are used exclusively for visible LEDs and electronic devices that are less susceptible to dislocations on device characteristics. ing.

  On the other hand, in LD and ultraviolet LED devices that require an epitaxial growth layer with a low dislocation density, there is a technique in which a substrate made of a single GaN material is used as a substrate for crystal growth, and a semiconductor multilayer film constituting an element portion is formed thereon. It is being considered. Hereinafter, such a GaN substrate for crystal growth is referred to as a GaN free-standing substrate.

  The GaN free-standing substrate is obtained by epitaxially growing a GaN layer with a reduced dislocation density on a heterogeneous substrate such as a sapphire substrate, peeling off the GaN layer from the substrate after growth, and using the obtained GaN layer as a self-standing GaN substrate. The method is common.

  For example, a GaN layer is formed on a sapphire substrate using a so-called ELO (Epitaxial Lateral Overgrowth) technique, which is a technique for obtaining a GaN layer with few dislocations by forming a mask having an opening on a base substrate and laterally growing from the opening. After that, it has been proposed to remove the sapphire substrate by etching or the like to obtain a GaN free-standing substrate (see, for example, Patent Document 1).

  Further, by growing GaN on a substrate such as sapphire through a TiN thin film having a network structure, the VAS method (Void-Assisted Separation) is used to form a void at the interface between the base substrate and the GaN layer. A technique for simultaneously enabling low dislocation is also disclosed (for example, see Non-Patent Document 1).

The GaN substrate obtained by these methods usually has a morphology such as pits and hillocks on its surface in an as-grown state, and it is difficult to grow an epi layer for device fabrication as it is. Therefore, the substrate surface is generally used after being polished to a mirror surface.
JP-A-11-251253 Y. Oshima et. Al., "Preparation of Freestanding GaN Wafers by Hydride Vapor Phase Epitaxy with Void-Assisted Separation.", Jpn. J. Appl. Phys. Vol. 42 (2003) pp. L1-L3

  In general, a semiconductor bulk crystal often has a structure in which multiple layers having different impurity concentrations are periodically stacked in the crystal growth direction (that is, the substrate thickness direction). In this multi-layer structure, the crystal is rotated during crystal growth, so that the crystal periodically passes through regions with temperature gradients and regions with different raw material and dopant concentrations. It is thought to appear by being laminated. The state of this multilayer structure is estimated from the fact that when the impurity concentration distribution is examined in the thickness direction of the substrate using SIMS or the like, the impurity concentration is accompanied by an amplitude of a fixed period determined by the rotation speed and growth rate of the crystal. it can.

  In addition, when the grown crystal is cut parallel to the growth direction and some kind of etching is performed on the cut surface, or the state of light emission is observed by applying excitation light, the "crystal growth interface" There is a clear striped pattern with a constant period parallel to the virtual crystal growth interface that should have existed at the time of growth. The observed stripes are called striations or growth stripes. That is, in the crystal, the shape of the crystal growth interface at the time when the specific region of the crystal has grown is engraved as impurity concentration unevenness = striation line. You can see the history. The period of the stripe pattern may be longer than the period of the impurity concentration distribution described above, and the two do not necessarily match, but the appearance of the stripe pattern as described above means that the thickness direction of the substrate This suggests that there are periodic fluctuations in the impurity concentration.

  Normally, the above-mentioned striations are formed in the compound semiconductor substrate such as GaAs or InP in the thickness direction of the substrate, but these crystals are cut out from a bulk crystal grown from a melt. Therefore, the difference in impurity concentration in the substrate thickness direction within the striation is relatively small, and it is presumed that this hardly has an adverse effect on device fabrication. This is because, as described above, the impurity concentration difference is taken into the crystal by periodically passing through a region having a temperature gradient or a region having different raw material or dopant concentration due to, for example, rotation of the crystal. However, in crystal growth of GaAs, InP, etc., the crystal growth interface is always in contact with the same melt surface, so the temperature gradient and dopant concentration gradient will be described later. This is because it is grown in an extremely small environment compared to the GaN crystal.

  On the other hand, since the GaN substrate is manufactured by vapor phase growth as described above, it is grown while periodically passing through an environment where the temperature gradient is steep compared to other compound semiconductor crystals grown from the melt. There are many cases. In addition, the crystal growth rate at the passing position of the crystal depends on the concentration of the raw material gas existing there, and the concentration of the impurity incorporated therein depends on the concentration of the dopant gas existing there. Since it is determined by the distribution of the raw material gas flow in the furnace, the variation tends to be very large as compared with crystals grown from the melt. For this reason, when the source gas is flowing in the furnace in a biased manner, if the crystal substrate is rotated, the appearance is averaged and crystals with uniform thickness and impurity concentration can be formed. Since it grows while alternately passing through large and small areas, it is often found that when it is examined in detail, it has a structure in which layers with extremely large impurity concentration differences are stacked in multiple layers. is there.

  The carrier concentration of a semiconductor substrate is generally measured by the van der Pauw method or the eddy current measurement method, and these are all viewed as an averaged carrier concentration as a bulk. Even if there is a periodic impurity concentration distribution, it cannot be detected. However, when there is a striation with a very large difference in impurity concentration as described above, it is equivalent to alternately stacking high carrier concentration layers and low carrier concentration layers. Even if the carrier concentration is high, if an electrode is attached to the substrate, ohmics cannot be obtained, or in devices that pass current perpendicular to the substrate, the operating voltage is increased, the heat generation of the device is increased, and the reliability is degraded. There was a problem that led to

  Therefore, the present invention has been made in view of the above circumstances, and the object of the present invention is to obtain the performance according to the value of the averaged electrical property obtained by measurement of a normal bulk crystal. It is an object to provide a group III-V nitride semiconductor substrate and a group III-V nitride light emitting device.

  Until now, it has not been performed by those skilled in the art to specify the carrier concentration distribution in the substrate thickness direction by paying attention to the striation of the substrate. However, as described above, the inventor of the present invention is that, even in a single crystal substrate having a uniform composition, impurity concentration distribution = carrier concentration distribution periodically exists in the thickness direction. If the substrate can be manufactured by paying attention and controlling the carrier concentration of the minimum carrier concentration layer in the substrate thickness direction to be equal to or higher than a specific value, devices generated due to electrode attachment and internal resistance of the substrate It was found that problems such as poor characteristics and poor reliability can be avoided.

  Further, when the crystal substrate is rotated during crystal growth, the present inventor has determined that the carrier concentration amplitude (minimum carrier concentration and maximum carrier concentration) in the striations on the substrate surface on the side closer to and far from the center of rotation. Focusing on the difference in concentration), even if the averaged carrier concentration looks uniform in the plane, if this amplitude is different, there will be a difference in the operating performance of the device fabricated on it. I found out.

The present invention has been made on the basis of the above findings in a new viewpoint that has not been focused on by those skilled in the art. That is, the group III-V nitride semiconductor substrate of the present invention is a self-supporting substrate made of a single crystal of a group III-V nitride semiconductor containing an n-type impurity, and is in the thickness direction of the substrate. The distribution of the n-type impurity concentration has periodic fluctuations, and the minimum value of the n-type impurity concentration is 5 × 10 ¹⁷ cm ⁻³ or more at an arbitrary position in the substrate surface. And

It is preferable that the amplitude of the periodic fluctuation of the n-type impurity concentration in the thickness direction of the substrate is 2 × 10 ¹⁸ cm ⁻³ or less at an arbitrary place in the substrate surface.

  The III-V nitride semiconductor is preferably hexagonal gallium nitride. In this case, the substrate surface is preferably a C-plane gallium surface. In addition, it is desirable to perform mirror polishing on the surface of the substrate.

  The n-type impurity may be silicon or oxygen. The diameter of the substrate is preferably 50 mm or more.

  Furthermore, an active layer made of at least a group III-V nitride semiconductor can be formed on the group III-V nitride semiconductor substrate to form a group III-V nitride light emitting device.

  According to the present invention, even if the average carrier concentration of the substrate is sufficiently high, the impurity concentration distribution in the depth direction varies periodically (that is, the low concentration layer and the high concentration layer have a periodic structure). Therefore, problems such as electrode ohmic failure, device characteristic failure, and reliability failure caused by the presence of a low concentration layer) can be avoided. As a result, the designed III-V nitride-based light emitting device can be manufactured with a high yield.

(Independent substrate)
The self-supporting substrate (self-supporting substrate) in the present invention refers to a substrate having a strength that can maintain its own shape and does not cause inconvenience in handling. In order to provide such strength, the thickness of the self-supporting substrate is preferably 200 μm or more. In consideration of easiness of cleavage after element formation, the thickness of the self-supporting substrate is preferably 1 mm or less. If the thickness is too thick, it will be difficult to cleave, and the cleaved surface will be uneven. As a result, when applied to, for example, a semiconductor laser, deterioration of device characteristics due to loss of reflection becomes a problem. Moreover, it is preferable that the diameter is 50 mm or more.

(III-V group nitride semiconductor)
As the III-V group nitride semiconductor in the present invention, a semiconductor represented by In _x Ga _y Al _1-xy N (0 ≦ x ≦ 1, 0 ≦ y ≦ 1, 0 ≦ x + y ≦ 1) is used. Can be mentioned. Of these, gallium nitride (GaN) is most preferred. This is because the properties required for the substrate material such as strength and manufacturing stability are satisfied. Further, the surface of the substrate is desirably a (0001) group III surface (gallium surface). This is because GaN-based crystals are more polar, the group III surface (gallium surface) is more chemically and thermally stable than the group V surface (nitrogen surface), and devices are easier to fabricate.

(Substrate surface)
In general, a large number of large irregularities such as hillocks and minute irregularities that appear to appear due to step bunching exist on the GaN-based epi surface of as-grown. These are not only factors that make the morphology, film thickness, composition, etc., non-uniform when the epi is grown thereon, but also the problem of reducing the exposure accuracy of the photolithography process in the device fabrication process. is there. Therefore, it is desirable that the substrate surface be polished as a flat mirror surface. However, a processing damage layer may remain on the surface of the polishing substrate due to polishing. The dislocation density is increased not only by dislocations generated by grown-in (during crystal growth) but also by dislocations introduced after crystal growth by polishing or the like, so that the dislocation density is lowered to obtain a substrate that is less prone to surface roughness. For this reason, it is desirable that the processing damage layer after polishing is removed by a technique such as wet etching, dry etching, or strain removal annealing. Needless to say, the surface of the substrate after the surface polishing is desirably flat, but the roughness is preferably 10 nm or less in terms of the arithmetic average roughness Ra measured in the 50 μm range. In addition, since the visible LED substrate does not require so much fine processing or the like, and cost competitiveness tends to be more important than that, a substrate (as-grown substrate) as obtained by crystal growth is used. It doesn't matter.

(Back side of substrate)
In general, a GaN-based free-standing substrate can be obtained by peeling off a hetero-growth substrate on a different kind of base substrate by some method. For this reason, the back surface of the substrate that has been peeled off is often roughened in a satin state, or a part of the base substrate is often attached. Further, it may not be flat due to warpage of the substrate. These cause the temperature distribution of the substrate to become non-uniform when growing the epi on the substrate, and as a result, the uniformity of the epi deteriorates or the reproducibility deteriorates. For this reason, it is generally desirable that the back surface of the substrate be polished flat.

(Substrate n-type impurities)
In addition to Si and O, the n-type impurity of the substrate can target Ge, Se, S, and the like.

(Carrier concentration)
The difference in impurity concentration in the substrate thickness direction is a problem in GaN substrates for light emitting devices such as LEDs and LDs that are operated by energization perpendicularly to the substrate. A carrier concentration of at least 5 × 10 ¹⁷ cm ⁻³ or more is required from the viewpoints of ease of attaching an electrode when manufacturing the substrate, reducing contact resistance between the electrode and the substrate, and reducing substrate resistance during energization. However, this carrier concentration is an apparent average carrier concentration, and so far, it has been performed by those skilled in the art to define the carrier concentration distribution in the substrate thickness direction by paying attention to the striation of the substrate. It wasn't. Even if the present inventor is a single crystal substrate having a uniform composition, the concentration distribution of impurities = carrier concentration distribution is periodically present in the thickness direction in the inside thereof, from the viewpoint of the substrate thickness direction. By producing the substrate by controlling the carrier concentration of the lowest carrier concentration layer to be 5 × 10 ¹⁷ cm ⁻³ or more, it is difficult to obtain device characteristics and reliability due to electrode attachment and internal resistance of the substrate. It was clarified that problems such as sexual defects could be avoided.

(Carrier concentration amplitude)
In addition, when the crystal substrate is rotated during crystal growth, the carrier concentration amplitude (the difference between the maximum value and the minimum value of the carrier concentration) in the striation in the substrate plane on the side closer to and far from the center of rotation. Will cause a difference. Even if the averaged carrier concentration looks uniform in the plane, if this amplitude is different, there will be a difference in the operating performance of the device fabricated thereon. The amplitude at an arbitrary location in the substrate surface is desirably 2 × 10 ¹⁸ cm ⁻³ or less. The reason for this range is that when the amplitude exceeds 2 × 10 ¹⁸ cm ⁻³ , the highest carrier concentration in the substrate surface exceeds 2.5 × 10 ¹⁸ cm ^−3, and the crystallinity of the substrate deteriorates. This is because the device performance starts to deteriorate.

(Measurement method of carrier concentration)
As a GaN substrate for a light emitting device, an n-type substrate doped with Si or O is usually used. Since the carrier concentration of the semiconductor substrate is often measured as averaged bulk information as described above, it is difficult to accurately measure the carrier concentration distribution in the striation. However, since the activation rate of these n-type dopants in the GaN crystal is close to 100%, the carrier concentration distribution in the striation is almost accurately grasped by measuring the dopant concentration distribution in the substrate thickness direction. can do. This dopant concentration can be easily measured by SIMS analysis that is generally used.

(Carrier concentration control method)
The carrier concentration of the lowest carrier concentration layer in the substrate thickness direction is set to 5 × 10 ¹⁷ cm ⁻³ or more, and the amplitude of the carrier concentration is set to 2 × 10 ¹⁸ cm ⁻³ or less at an arbitrary position in the plane. In order to fabricate the substrate in such a manner, the temperature distribution in the crystal growth furnace is made uniform and the flow of the source gas and dopant gas is made uniform, and then an appropriate crystal growth rate and crystal rotation speed are set. In combination, the optimum condition is selected, and it is confirmed by the SIMS analysis whether the target impurity concentration distribution is satisfied, and the growth condition is finally determined. The individual crystal growth conditions differ depending on the furnace used, and since there are many optimum conditions depending on the combination of the parameters, they cannot be uniquely determined.

(Substrate manufacturing method)
The group III-V nitride semiconductor substrate of the present invention can be obtained by growing a single crystal of a group III-V nitride semiconductor on a different substrate and then peeling it. A single crystal of a III-V nitride semiconductor is desirably grown by HVPE (hydride vapor phase epitaxy). This is because the HVPE method has a high crystal growth rate and is suitable for manufacturing a substrate that requires thick film growth. Hereinafter, the HVPE reactor used when manufacturing a GaN free-standing substrate will be described in detail.

(HVPE reactor)
FIG. 1 shows a configuration of a general HVPE furnace used for crystal growth of a GaN free-standing substrate.
This HVPE reaction furnace 10 is a hot wall type that is heated outside a horizontally long quartz reaction tube 1 by a heater comprising a raw material heating heater 2a and a crystal growth region heating heater 2b. On the upstream side), an ammonia gas introduction pipe 3 for introducing NH ₃ gas as a group V source, a hydrochloric acid gas introduction pipe 4 for introducing hydrochloric acid (HCl) gas for forming GaCl as a group III source, And a doping gas introduction pipe 5 for introducing a dopant gas for controlling conductivity. Further, the hydrochloric acid gas introduction pipe 4 is expanded in the middle to form a gallium melt reservoir 6 and can accommodate a gallium melt 7 in which metal gallium is melted. On the other hand, in the vicinity of the center of the drawing in the quartz reaction tube 1, a base substrate 8 is disposed, and a substrate folder 9 provided so as to be rotatable about a rotation shaft 9a is disposed. Furthermore, an exhaust pipe 11 for releasing exhaust gas to the outside is provided on the right side (downstream side) of the drawing.

  In order to grow GaN using this HVPE reactor 10, first, the raw material heating heater 2a is heated to 800 ° C., the crystal growth region heating heater 2b is heated to 1000 ° C., and the gallium melt reservoir 6 is made of Ga. The metal gallium is melted by heating to a temperature higher than the melting point to form a gallium melt 7.

Next, NH ₃ gas which is a group V raw material is introduced from the ammonia gas introduction pipe 3, HCl gas for generating a group III raw material is introduced from the hydrochloric acid gas introduction pipe 4, and a gas containing a dopant component is introduced from the doping gas introduction pipe 5. To do. From the viewpoint of controlling the reactivity, HCl gas and NH ₃ gas, which are raw material gases, are used by being mixed with a carrier gas such as H ₂ gas. In addition, since HVPE growth involves a considerable amount of Si mixing from the quartz member in the furnace, depending on the growth conditions, a GaN crystal containing n-type impurities can be grown without flowing a doping gas.

In the hydrochloric acid gas introduction pipe 4, HCl gas comes into contact with the gallium melt 7 along the way,
A reaction of Ga + HCl → GaCl + (1/2) H ₂ occurs to generate gallium chloride GaCl.

The mixed gas of GaCl gas and carrier gas H _{2 and} the mixed gas of ammonia NH ₃ and carrier gas H ₂ are carried in the space in the reaction tube 1 in the direction of the arrow, and on the base substrate 8 provided in the substrate folder 9. Thus, a reaction of GaCl + NH ₃ → GaN + HCl + H ₂ occurs, and GaN is deposited on the base substrate 8. Note that the base substrate 8 as a base for crystal growth is fixed to the substrate folder 9 supported by the rotation shaft 9a, and is rotated during the growth. The gas introduced into the reaction tube is guided to a detoxification facility by the downstream exhaust pipe 11, subjected to a detoxification process, and then discharged to the atmosphere.

(Example of improved HVPE furnace)
The HVPE furnace can be improved to a structure as shown in FIG. The HVPE furnace 20 is formed in the same manner as the HVPE furnace shown in FIG. 1 except that the substrate folder 19 provided so as to be rotatable about the rotating shaft 19a is inclined by 10 ° with respect to the flow of the source gas. ing.
By comprising in this way, a flow-path cross-sectional area becomes narrow and the flow velocity becomes large as the flow of source gas goes downstream. By this action, the influence of the decrease in the raw material concentration due to the exhaustion of the GaCl raw material on the upstream side is offset, and the growth rate is suppressed from decreasing from the upstream to the downstream of the flow.

(Modification of HVPE furnace)
The HVPE furnace can be modified to have a structure as shown in FIG. This HVPE furnace 30 is shown in FIG. 1 except that a large number of substrate folders 29 provided so as to be rotatable about a rotating shaft 29a are of a simultaneous growth type and three substrates having a diameter of 2 inches can be placed. It is formed like the HVPE furnace.
The substrate folder 29 can be rotated around the rotation axis 29a, but the individual base substrate 8 itself does not rotate. That is, the point that the base substrate 8 revolves without rotating during HVPE growth is different from the HVPE furnace shown in FIG.

(Peeling method from the base substrate)
As a method of peeling a GaN-based semiconductor single crystal from a base substrate after growing it, a void formation peeling method (VAS method) can be used. The VAS method is excellent in that a large-diameter substrate can be peeled off with good reproducibility and a GaN free-standing substrate with low dislocations and uniform characteristics can be obtained. Hereinafter, the VAS method will be described with reference to FIG.

(VAS method)
In the VAS method, a GaN layer 43 (for example, a Si-doped GaN layer 43) is first grown on a heterogeneous substrate 41 such as sapphire by a MOVPE method through a low-temperature growth GaN buffer layer (not shown) by about 0.5 μm ( a). Next, a metal Ti thin film 45 is deposited on the GaN layer 43 to a thickness of about 20 nm (b), and the substrate is placed in an electric furnace and heated to about 1050 ° C. in a hydrogen gas stream containing ammonia. Apply heat treatment. By doing so, a part of the GaN layer 43 is etched to generate a void layer 46 having high-density voids, and the Ti thin film 45 is nitrided to form high-density submicron holes on the surface. It changes into the network-like TiN layer 47 formed in (c). This substrate is placed in an HVPE reactor as shown in FIGS. 1 to 3 as a base substrate, and a thick GaN layer 48 is grown thereon (d). In the process of cooling the HVPE reactor, the thick GaN layer 48 is naturally separated from the underlying substrate with the void layer 46 as a boundary, and a GaN free-standing substrate 49 is obtained (e). By planarizing the front and back surfaces of the obtained substrate, a GaN free-standing substrate 50 is obtained (f).

(GaN light emitting device)
The obtained GaN free-standing substrate is suitable for use in manufacturing a light-emitting diode by epitaxially growing a group III-V nitride semiconductor crystal on the GaN substrate by MOVPE.

  The present invention will be described in more detail with reference to the following examples, but the present invention is not limited thereto.

(Preparation of base substrate by VAS method)
First, the base substrate (void formation GaN template) used in the comparative examples and examples described below was manufactured by the VAS method.

As shown in FIG. 4, a Si-doped GaN layer 43 is formed on a φ2 inch diameter sapphire C-plane just substrate (heterogeneous substrate 41) by a MOVPE method through a 20 nm low-temperature growth GaN buffer layer (not shown). .5 μm growth (a). The growth pressure was normal pressure, the substrate temperature during buffer layer growth was 600 ° C., and the substrate temperature during epi layer growth was 1100 ° C. As raw materials, trimethylgallium (TMG) was used as a group III raw material, ammonia was used as a group V raw material, and monosilane was used as a dopant. The carrier gas is a mixed gas of hydrogen and nitrogen. The crystal growth rate was 4 μm / h. The carrier concentration of the epi layer was 2 × 10 ¹⁸ cm ⁻³ .

  Next, a metal Ti thin film 45 was deposited on the GaN epi substrate to a thickness of 20 nm (b). The substrate thus obtained was put in an electric furnace and heat-treated at 1050 ° C. for 20 minutes in a hydrogen gas stream containing 20% ammonia. As a result, a part of the Si-doped GaN layer 43 is etched to generate a high-density void layer (void layer 46), and the Ti thin film 45 is nitrided to form submicron fine holes at a high density on the surface. The resulting titanium nitride layer (TiN layer 47) was changed to (c). The comparative examples and examples to be described below are all performed using this base substrate (void-formed GaN template).

[Comparative Example 1]
(When the minimum value of the impurity concentration is not 5 × 10 ¹⁷ cm ⁻³ or more and the in-plane amplitude is not 2 × 10 ¹⁸ cm ⁻³ or less at a location 20 mm away from the center of the substrate)
First, using the HVPE furnace shown in FIG. 1, a thick GaN film was grown on the void-formed GaN template (underlying substrate 8). The growth condition of HVPE is that a GaN layer is formed to a target thickness by using a supply gas containing a source gas composed of 8 × 10 ⁻³ atm gallium chloride and 4.8 × 10 ⁻² atm ammonia in a carrier gas. Grown up. As the carrier gas, nitrogen gas containing 5% of hydrogen was used. The growth conditions for the GaN layer were atmospheric pressure and a substrate temperature of 1000 ° C. Also, in the GaN crystal growth process, silicon was doped by supplying dichlorosilane as a doping source gas to the substrate region.

  5 and 6 are diagrams showing the film thickness distribution of GaN crystals grown using the HVPE furnace shown in FIG.

  FIG. 5 shows the growth with intentionally stopping the substrate rotation in order to clarify the growth rate distribution with respect to the flow direction of the source gas. Under this growth condition, a sufficiently large amount of ammonia gas is allowed to flow with respect to the molar flow rate of gallium chloride. Therefore, the growth rate of GaN is limited to supply of gallium chloride. Since gallium chloride is consumed by crystal growth of GaN, the concentration of the raw material decreases as it goes downstream, and thus the growth rate also slows down. From FIG. 5, it can be seen that under the growth conditions, the growth rate is reduced to about 1/5 during the substrate diameter of 50 mm.

  FIG. 6 is a film thickness distribution of a GaN crystal grown by rotating the substrate at 10 rpm under the same growth conditions. Due to the rotation of the substrate, the difference in growth rate upstream and downstream of the source gas flow was averaged, and the surface was considerably flattened.

  Under the above growth conditions, the substrate is rotated at 10 rpm to form a GaN layer 48 having a center film thickness of about 600 μm (FIG. 4D), and after the substrate is peeled to form a freestanding substrate 49 (FIG. 4E). The front and back surfaces of this substrate were lapped and mirror-polished, respectively, to obtain a 430 μm-thick GaN free-standing substrate 50 (FIG. 4F).

Nine 5 mm square samples were cut out in a line along the diameter of the substrate, and each carrier was measured for the carrier concentration by the van der Pauw method with In electrodes attached. The value was 1.1 ± 0.1 × 10 ^18. It was in the range of cm ⁻³ . Next, the Si concentration distribution in the thickness direction of the substrate was measured with respect to the sample cut out from the center of the substrate and 10 mm and 20 mm away from the center. FIGS. 7 to 9 show the SIMS measurement results at the center of the substrate and at locations 10 mm and 20 mm away from the center, respectively.

FIG. 7 shows a Si concentration profile at the center of the GaN free-standing substrate. The depth distribution of the Si concentration measured with the sample cut from the center of the substrate does not vary so much, and the maximum value is 1.23 × 10 ¹⁸ cm ⁻³ and the minimum value is 9. 96 × ¹⁰ 17 ^{cm -3,} and an average value 1.12 × ¹⁰ 18 ^{cm -3.} Here, assuming that the amplitude of fluctuation of the Si concentration is (maximum concentration−minimum concentration), the value is 2.34 × 10 ¹⁷ cm ⁻³ .

FIG. 8 shows a Si concentration profile at a point 10 mm from the center of the GaN free-standing substrate. The depth distribution of the Si concentration measured with a sample cut from a location 10 mm away from the center of the substrate is slightly larger than that of the central portion, with a maximum value of 1.75 × 10 ¹⁸ cm ⁻³ and a minimum value of 5. 58 × ¹⁰ 17 ^{cm -3,} and an average value 1.12 × ¹⁰ 18 ^{cm -3.} Accordingly, the amplitude of the Si concentration at this measurement point is 1.19 × 10 ¹⁸ cm ⁻³ .

FIG. 9 shows a Si concentration profile at a point 20 mm from the center of the GaN free-standing substrate. The depth distribution of the Si concentration measured with a sample cut from a location 20 mm away from the center of the substrate has a larger variation than that of the central portion, with a maximum value of 3.21 × 10 ¹⁸ cm ⁻³ and a minimum value of 3. 23 × ¹⁰ 17 ^{cm -3,} and an average value 1.14 × ¹⁰ 18 ^{cm -3.} Thus, the amplitude of the Si concentration at a location 20 mm away from the center of the substrate is 2.89 × 10 ¹⁸ cm ⁻³ .

Therefore, it can be said that this substrate has an Si concentration amplitude of about 2.89 × 10 ¹⁸ cm ⁻³ or less at an arbitrary position in the plane. In SIMS measurement, an abnormal impurity concentration may be measured on the outermost surface of the substrate due to an adsorbed impurity or the like. In addition, even when measured on a specific point where a crystal defect or the like exists, the impurity concentration may be different from that of the parent phase portion that occupies most of the substrate. For this reason, in this specification, the value measured except these special points shall be handled. Therefore, the above-described impurity concentration values are handled excluding high concentration measurement values that are clearly judged to be the influence of the substrate surface. The same applies to the embodiments described below.

An LED epi having a structure shown in FIG. 10 was grown on a GaN free-standing substrate manufactured by the same method as the above-described substrate, and the variation in the LED characteristics was investigated by forming a chip. The LED structure epi was grown using a MOVPE growth apparatus. The raw materials used for the growth are TMG, TMA (trimethylaluminum), TMI (trimethylindium), and NH ₃ . In the growth, first, the substrate is heated to 1150 ° C. in a mixed gas flow of NH ₃ : hydrogen = 1: 1, held for 5 minutes after the temperature is stabilized, and then the group III raw material necessary for the growth of the first layer is obtained. The procedure was to start flowing. The grown epi structure has an Si-doped n-type GaN layer 52 of 1 μm and an In _0.15 Ga _0.85 N / GaN-3-MQW active layer 53 (well layer 3 nm, barrier layer 10 nm in order from the GaN substrate 51 side. ), The Mg-doped p-type Al _0.1 Ga _0.9 N layer 54 is 40 nm, and the p-type GaN layer 55 is 500 nm. The growth of the MQW layer was performed by lowering the growth temperature to 800 ° C. The growth temperature of the other layers is 1150 ° C. The growth pressure was all normal pressure. After the epi layer growth, an n-type electrode 56 made of Ti / Al was formed on the back surface of the GaN substrate 51, and a p-type electrode 57 made of Ni / Au was formed on the surface of the p-type GaN layer 55.

  The drive voltage Vf at 20 mA of the manufactured LED chip varies greatly from 3.25 to 6.02 V within the substrate surface, and the Vf tends to increase as the chip obtained on the outer peripheral side of the substrate. Assuming Vf <3.5V as an acceptable product, the acceptance rate for this substrate was 45%. Further, even in the result of the reliability test performed by the constant current drive, there was a tendency that the fluctuation range of the light emission output was larger as the chip obtained on the outer peripheral side of the substrate.

[Example 1]
(When the minimum value of the impurity concentration is 5 × 10 ¹⁷ cm ⁻³ or more and the in-plane amplitude is 2 × 10 ¹⁸ cm ⁻³ or less)
A GaN free-standing substrate was fabricated under the same conditions as in Comparative Example 1 except that the HVPE reactor having the structure shown in FIG.

  11 and 12 are diagrams showing the film thickness distribution of GaN crystals grown using the HVPE furnace shown in FIG.

FIG. 11 shows growth with intentionally stopping the substrate rotation in order to clarify the growth rate distribution with respect to the flow direction of the source gas. Gallium chloride is consumed in the crystal growth of GaN, and the concentration of the raw material decreases as it goes downstream. However, since the substrate is inclined as described above, the growth rate tends to be slower on the downstream side than Comparative Example 1. Can be greatly eased.
FIG. 12 shows the film thickness distribution of a GaN crystal grown by rotating the substrate at 10 rpm under the same growth conditions. Due to the rotation of the substrate, the difference in growth rate between the upstream and downstream of the source gas flow was averaged, and the surface was almost flattened.

A GaN free-standing substrate was produced using the HVPE furnace shown in FIG. 2 under the same conditions as in Comparative Example 1. Nine 5 mm square samples were cut out in a row along the diameter of the substrate, and each was attached with an In electrode and van. The carrier concentration was measured by the der Pauw method. As a result, the value of carrier concentration was within the range of 1.2 ± 0.1 × 10 ¹⁸ cm ⁻³ . Next, the Si concentration distribution in the thickness direction of the substrate was measured with respect to the sample cut out from the center of the substrate and 10 mm and 20 mm away from the center. FIGS. 13 to 15 show the SIMS measurement results at the center of the substrate and at locations 10 mm and 20 mm away from the center, respectively.

FIG. 13 shows the Si concentration profile at the center of the GaN free-standing substrate. The depth distribution of the Si concentration measured with the sample cut out from the center of the substrate is a maximum value of 1.31 × 10 ¹⁸ cm ⁻³ and a minimum value of 1.10 × 10 ¹⁸ cm ^{− when} measured to a depth of about 10 μm. ³ and the average value was 1.21 × 10 ¹⁸ cm ⁻³ . Here, assuming that the amplitude of fluctuation of the Si concentration is (maximum concentration−minimum concentration), the value is 2.10 × 10 ¹⁷ cm ⁻³ .

FIG. 14 shows the Si concentration profile at a point 10 mm from the center of the GaN free-standing substrate. The depth distribution of the Si concentration measured with a sample cut from a location 10 mm away from the center of the substrate has a maximum value of 1.40 × 10 ¹⁸ cm ⁻³ , a minimum value of 1.07 × 10 ¹⁸ cm ⁻³ , and an average value of 1 20 × 10 ¹⁸ cm ⁻³ . Thus, the amplitude of the Si concentration at a location 10 mm away from the center of the substrate is 3.3 × 10 ¹⁷ cm ⁻³ .

FIG. 15 shows a Si concentration profile at a point 20 mm from the center of the GaN free-standing substrate. The depth distribution of the Si concentration measured with a sample cut from a location 20 mm away from the center of the substrate is almost the same as that of the central portion, with a maximum value of 1.45 × 10 ¹⁸ cm ⁻³ and a minimum value of 8.58 × 10. ^{It was 17} cm ⁻³ and the average value was 1.17 × 10 ¹⁸ cm ⁻³ . Thus, the amplitude of the Si concentration is 5.92 × 10 ¹⁷ cm ⁻³ .

From the above results, it can be seen that the amplitude of the Si concentration tends to gradually increase toward the outer peripheral portion of the substrate, but the amplitude of the Si concentration at the center of the substrate is 2.10 × 10 ¹⁷ cm ^−3, which is 20 mm from the center. That is, the amplitude of the Si concentration measured at a point 5 mm from the outermost periphery of the substrate was 5.92 × 10 ¹⁷ cm ⁻³ . It is estimated that the amplitude will be in the 10 ¹⁷ cm ⁻³ range.

  On the GaN free-standing substrate, an LED epi having the structure shown in FIG. 10 was grown as in Comparative Example 1, and the chip was made into chips to investigate in-plane variation in LED characteristics. The drive voltage Vf at 20 mA of the manufactured LED chip was relatively uniform at 3.25 to 3.55 V within the substrate surface, and a substantially uniform distribution of Vf was observed over the entire surface of the substrate. Assuming that Vf <3.5 V is an acceptable product, the acceptance rate of this substrate was 98%. Further, even in the result of the reliability test performed by constant current driving, no distribution corresponding to the chip acquisition position was found in the plane of the substrate.

[Example 2]
(When the minimum value of the impurity concentration is 5 × 10 ¹⁷ cm ⁻³ or more but the in-plane amplitude is not 2 × 10 ¹⁸ cm ⁻³ or less)
Using the HVPE apparatus shown in FIG. 1, a GaN free-standing substrate was fabricated under the same conditions as in Comparative Example 1. The difference from Comparative Example 1 is that the flow rate of dichlorosilane, which is a dopant gas, is increased so that the minimum value of the Si concentration does not become lower than 5 × 10 ¹⁷ cm ⁻³ even on the outer peripheral side of the crystal. As a result, when the surface of the crystal after the growth was observed, fine irregularities were observed particularly on the outer peripheral side of the crystal. The same polishing process as in Comparative Example 1 was applied to the front and back surfaces of this crystal, and nine 5 mm square samples were cut out in a line along the diameter from the obtained substrate, each with an In electrode, and a carrier by van der Pauw method. Concentration was measured. The carrier concentration was slightly higher than that in Comparative Example 1, and was in the range of 1.6 ± 0.1 × 10 ¹⁸ cm ⁻³ . Next, the Si concentration distribution in the thickness direction of the substrate was measured with respect to the sample cut out from the center of the substrate and 10 mm and 20 mm away from the center.

The depth distribution of the Si concentration measured with the sample cut out from the center of the substrate does not change so much, and the maximum value is 1.71 × 10 ¹⁸ cm ⁻³ and the minimum value is 1 when measured to a depth of about 6 μm. It was .48 × 10 ¹⁸ cm ⁻³ and the average value was 1.61 × 10 ¹⁸ cm ⁻³ . Therefore, the amplitude of variation of the Si concentration is 2.30 × 10 ¹⁷ cm ⁻³ .

The depth distribution of the Si concentration measured with a sample cut from a location 10 mm away from the center of the substrate is slightly larger than that of the central portion, with a maximum value of 2.19 × 10 ¹⁸ cm ⁻³ and a minimum value of 1. It was 08 × 10 ¹⁸ cm ⁻³ and the average value was 1.63 × 10 ¹⁸ cm ⁻³ . Thus, the amplitude of the Si concentration at this measurement point is 1.11 × 10 ¹⁸ cm ⁻³ .

The depth distribution of the Si concentration measured with a sample cut from a location 20 mm away from the center of the substrate has a larger variation than that of the central portion, with a maximum value of 3.61 × 10 ¹⁸ cm ⁻³ and a minimum value of 8. 25 × ¹⁰ 17 ^{cm -3,} and an average value 1.60 × ¹⁰ 18 ^{cm -3.} Thus, the amplitude of the Si concentration at a location 20 mm away from the center of the substrate is 2.79 × 10 ¹⁸ cm ⁻³ . Therefore, it can be said that this substrate has an Si concentration amplitude of about 2.79 × 10 ¹⁸ cm ⁻³ or less at an arbitrary position in the plane.

  On the GaN free-standing substrate, an LED epi having the structure shown in FIG. 10 was grown as in Comparative Example 1, and the chip was made into chips to investigate in-plane variation in LED characteristics. The drive voltage Vf at 20 mA of the manufactured LED chip was relatively uniform from 3.35 to 3.81 V within the substrate surface, and a substantially uniform distribution of Vf was observed over the entire surface of the substrate. Assuming that Vf <3.5V is an acceptable product, this substrate had a passing rate of 91%. However, the light emission output of the LED tends to be low at the outer peripheral portion of the substrate, and in the result of the reliability test performed by constant current driving, a tendency that the fluctuation of the light output is larger as the outer peripheral portion is observed.

[Example 3]
(When the minimum value of the impurity concentration is 5 × 10 ¹⁷ cm ⁻³ or more and the in-plane amplitude is 2 × 10 ¹⁸ cm ⁻³ or less)
Using the HVPE apparatus shown in FIG. 2, a GaN free-standing substrate was fabricated under the same conditions as in Example 1. The difference from the first embodiment is that the tilt angle of the substrate of the HVPE apparatus is reduced to 5 °. Nine 5 mm square samples were cut out in a line along the diameter from the obtained substrate, and each carrier concentration measured by the van der Pauw method with an In electrode attached was almost the same as in Example 1, 1.5 ± It was in the range of 0.2 × 10 ¹⁸ cm ⁻³ . Next, the Si concentration distribution in the thickness direction of the substrate was measured with respect to the sample cut out from the center of the substrate and 10 mm and 20 mm away from the center.

The depth distribution of the Si concentration measured with the sample cut out from the center of the substrate is a maximum value of 1.69 × 10 ¹⁸ cm ⁻³ and a minimum value of 1.28 × 10 ¹⁸ cm ^{− when} measured to a depth of about 10 μm. ³ and the average value was 1.49 × 10 ¹⁸ cm ⁻³ . Here, assuming that the amplitude of fluctuation of the Si concentration is (maximum concentration−minimum concentration), the value is 4.1 × 10 ¹⁷ cm ⁻³ .

The depth distribution of the Si concentration measured with a sample cut from a location 10 mm away from the center of the substrate has a maximum value of 2.16 × 10 ¹⁸ cm ⁻³ , a minimum value of 8.62 × 10 ¹⁷ cm ⁻³ , and an average value of 1 It was 51 × 10 ¹⁸ cm ⁻³ . Thus, the amplitude of the Si concentration at a location 10 mm away from the center of the substrate is 1.30 × 10 ¹⁸ cm ⁻³ .

The depth distribution of the Si concentration measured with a sample cut from a location 20 mm away from the center of the substrate has a maximum value of 2.44 × 10 ¹⁸ cm ⁻³ , a minimum value of 5.43 × 10 ¹⁷ cm ⁻³ , and an average value of 1 It was 51 × 10 ¹⁸ cm ⁻³ . Accordingly, the amplitude of the Si concentration is 1.90 × 10 ¹⁸ cm ⁻³ .

From the above results, it can be seen that the amplitude of the Si concentration tends to gradually increase toward the outer peripheral portion of the substrate, and in this substrate, the amplitude of the Si concentration is within 2 × 10 ¹⁸ cm ⁻³ within the substrate plane. It is estimated that it has a certain degree of distribution.

  On the GaN free-standing substrate, an LED epi having the structure shown in FIG. 10 was grown as in Comparative Example 1, and the chip was made into chips to investigate in-plane variation in LED characteristics. The drive voltage Vf at 20 mA of the manufactured LED chip was relatively uniform from 3.30 to 3.77 V within the substrate surface, and a substantially uniform distribution of Vf was observed over the entire surface of the substrate. Assuming Vf <3.5V as an acceptable product, this substrate had a acceptance rate of 92%. Even in this substrate, the LED light emission output tends to be slightly low in the outer peripheral portion of the substrate, and even in the result of the reliability test performed by constant current driving, a tendency that the light emission output fluctuates greatly in the outer peripheral portion was observed. However, no abnormal behavior was observed.

[Comparative Example 2]
(When the minimum value of the impurity concentration is not 5 × 10 ¹⁷ cm ⁻³ or more and the in-plane amplitude is not 2 × 10 ¹⁸ cm ⁻³ or less)
A GaN free-standing substrate was manufactured under the same conditions as in Comparative Example 1 using an HVPE apparatus having a susceptor of a multiple-growth type as shown in FIG.
Nine 5 mm square samples were cut out in a line along the diameter from the obtained substrate, and each carrier concentration measured by the van der Pauw method with an In electrode attached was almost the same as in Comparative Example 1. It was in the range of 1 ± 0.2 × 10 ¹⁸ cm ⁻³ . Next, the Si concentration distribution in the thickness direction of the substrate was measured with respect to the sample cut out from the center of the substrate and 10 mm and 20 mm away from the center.

In this substrate, a large variation was observed in the depth distribution of the Si concentration of the sample cut from any position from the center to the outer periphery of the substrate. In the center of the substrate, the measurement was performed up to a depth of about 6 μm. The maximum value was 3.09 × 10 ¹⁸ cm ⁻³ , the minimum value was 3.89 × 10 ¹⁷ cm ⁻³ , and the average value was 1.13 × 10 ¹⁸ cm ⁻³ . Therefore, the amplitude of variation of the Si concentration is 2.70 × 10 ¹⁸ cm ⁻³ .

The depth distribution of the Si concentration measured with a sample cut from a location 10 mm away from the center of the substrate has a maximum value of 3.03 × 10 ¹⁸ cm ⁻³ , a minimum value of 4.10 × 10 ¹⁷ cm ⁻³ , and an average value of 1 It was 13 × 10 ¹⁸ cm ⁻³ . Accordingly, the amplitude of the Si concentration at this measurement point is 2.62 × 10 ¹⁸ cm ⁻³ .

The depth distribution of the Si concentration measured with a sample cut from a location 20 mm away from the center of the substrate has a maximum value of 3.21 × 10 ¹⁸ cm ⁻³ , a minimum value of 3.35 × 10 ¹⁷ cm ⁻³ , and an average value of 1 It was 15 × 10 ¹⁸ cm ⁻³ . Thus, the amplitude of the Si concentration at a location 20 mm away from the center of the substrate is 2.87 × 10 ¹⁸ cm ⁻³ . Therefore, it can be said that this substrate has an Si concentration amplitude of about 2.87 × 10 ¹⁸ cm ⁻³ or less at an arbitrary position in the plane.

  On the GaN free-standing substrate, an LED epi having the structure shown in FIG. 10 was grown as in Comparative Example 1, and the chip was made into chips to investigate in-plane variation in LED characteristics. The drive voltage Vf at 20 mA of the manufactured LED chip varied greatly from 3.31 to 5.83 V within the substrate surface, and it was observed that the average Vf was high on the entire surface of the substrate. Assuming Vf <3.5V as an acceptable product, this substrate had a pass rate of only 15%. In addition, even in the result of the reliability test performed by constant current driving, no distribution corresponding to the chip acquisition position was observed in the plane of the substrate, and a tendency that the fluctuation of the light emission output was large overall was observed.

  Although the present invention has been described based on the embodiments, this is an exemplification, and various modifications can be made to combinations of various processes, and such modifications are also within the scope of the present invention. That is understood by the contractor. For example, in the embodiment, by using an HVPE reactor (FIG. 2) obtained by improving the conventional HVPE reactor (FIG. 1), the way in which impurities are incorporated into the crystal is prevented from periodically fluctuating. However, even in the conventional HVPE reactor shown in FIG. 1, for example, if the raw material flow rate is extremely increased, the utilization efficiency of the raw material is sacrificed, but the difference in the growth rate between the upstream and downstream of the flow is reduced, It is possible to make the distribution of the impurity concentration in the depth direction uniform. In short, it is important to reduce the fluctuation range of the impurity concentration in the depth direction of the substrate, and any means can be used.

It is a cross-sectional schematic diagram which shows the structure of the conventional HVPE reactor used for crystal growth. It is a cross-sectional schematic diagram which shows the structure which improved the HVPE reactor of FIG. It is a cross-sectional schematic diagram which shows the structure which deform | transformed the HVPE reactor of FIG. It is a schematic diagram for demonstrating the manufacturing process (VAS method) of a GaN substrate. In Comparative Example 1, it is a graph showing the film thickness distribution of GaN when growing with the substrate rotation stopped. In Comparative Example 1, it is a graph showing the film thickness distribution of GaN when growing by rotating the substrate. FIG. 4 is a diagram showing a Si concentration profile in the central portion of a GaN free-standing substrate crystal produced in Comparative Example 1. 6 is a diagram showing a Si concentration profile at a point 10 mm from the center of the GaN free-standing substrate crystal produced in Comparative Example 1. FIG. FIG. 6 is a diagram showing a Si concentration profile at a point 20 mm from the center of the GaN free-standing substrate crystal produced in Comparative Example 1. It is sectional drawing which shows the structure of LED produced by the comparative example and the Example. In Example 1, it is a graph which shows the film thickness distribution of GaN when it stops growing a substrate and it grew. In Example 1, it is a graph which shows the film thickness distribution of GaN when it grows by rotating a board | substrate. FIG. 3 is a diagram showing a Si concentration profile in the central portion of a GaN free-standing substrate crystal produced in Example 1. FIG. 3 is a diagram showing a Si concentration profile at a point 10 mm from the center of the GaN free-standing substrate crystal produced in Example 1. FIG. 3 is a view showing a Si concentration profile at a point 20 mm from the center of the GaN free-standing substrate crystal produced in Example 1.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Reaction tube 2a Heater for raw material heating 2b Heater for crystal growth region heating 3 Ammonia gas introduction pipe 4 Hydrochloric acid gas introduction pipe 5 Doping gas introduction pipe 6 Gallium melt reservoir 7 Gallium melt 8 Base substrate 9, 19, 29 Substrate folder 9a , 19a, 29a Rotating shaft 10, 20, 30 HVPE reactor 41 Heterogeneous substrate 43 Si-doped GaN layer 45 Ti thin film 46 Void layer 47 TiN layer 48 GaN layer 49 GaN free-standing substrate 50 GaN free-standing substrate 51 GaN substrate 52 Si-doped n-type GaN layer 53 InGaN / GaN-MQW active layer 54 Mg-doped p-type Al _0.10 GaN layer 55 p-type GaN layer 56 n-type electrode (Ti / Al)
57 p-type electrode (Ni / Au)
60 LED chip

Claims (9)

A self-supporting substrate made of a single crystal of a group III-V nitride-based semiconductor containing an n-type impurity, wherein the distribution of the n-type impurity concentration has a periodic variation with respect to the thickness direction of the substrate. And a minimum value of the n-type impurity concentration is 5 × 10 ¹⁷ cm ⁻³ or more at an arbitrary location in the substrate surface. 2. The III according to claim 1, wherein the amplitude of the periodic variation of the n-type impurity concentration in the thickness direction of the substrate is 2 × 10 ¹⁸ cm ⁻³ or less at an arbitrary location in the substrate surface. -Group V nitride semiconductor substrate.   2. The group III-V nitride semiconductor substrate according to claim 1, wherein the group III-V nitride semiconductor is hexagonal gallium nitride.   4. The III-V nitride semiconductor substrate according to claim 3, wherein the substrate surface is a C-plane gallium surface.   2. The group III-V nitride semiconductor substrate according to claim 1, wherein the substrate surface is mirror-polished.   2. The group III-V nitride semiconductor substrate according to claim 1, wherein the n-type impurity is silicon.   2. The group III-V nitride semiconductor substrate according to claim 1, wherein the n-type impurity is oxygen.   2. The group III-V nitride semiconductor substrate according to claim 1, wherein the substrate has a diameter of 50 mm or more.   9. A group III-V nitride comprising an active layer made of at least a group III-V nitride semiconductor formed on the group III-V nitride semiconductor substrate according to any one of claims 1 to 8. Physical light emitting device.

Images